CH329210A - Method for detecting and neutralizing an incipient explosion and apparatus for carrying out the method - Google Patents

Description

Method for detecting and neutralizing an incipient explosion and apparatus for implementing the method The present invention comprises a method for detecting and neutralizing incipient explosions and an apparatus for implementing this method.

<B>11</B> exists a method for detecting the onset of an explosion and for suppressing it before Wait could develop, whereas previously an was limited to limiting or deflecting the explosive waves instead to ensure the complete suppression of an incipient explosion before Von Watte reaches Pressures which could cause damage.

The present invention uses the same general principle to detect an explosion and suppress it before it can develop and this in the case of applications where it is more important to detect the illumination produced by the initial flame of the explosion or by the ignition source of the explosion, i.e. an incendiary bullet or a spark.

In the case of an installation for the production and handling of dust remaining in suspension in Jans Fair and of such a nature that it could give rise to an explosion, it has already been proposed to associate an explosive charge with each conduit which connects neighboring chambers of the installation, together with means for detonating this charge by the flame of the explosion occurring in the dust charge, to prevent the propagation of an explosion in the entire installation or in the neighboring parts thereof. A photoelectric cell may be arranged in side tubes issuing from a conduit in such a way that the light emitted by the flame of the dust explosion acts on the photoelectric cell to detonate the charge.

Such an arrangement is only capable of mitigating the effect of an explosion by preventing it from spreading from one Part of an installation to neighboring Parts which are connected to it by pipes. Nothing is done to protect the Part of the installation in which the explosion originates and it is the flame resulting from an actual explosion that the cell must detect.

Thus, the prior method that has just been recalled is limited to limiting the explosion to the container where it originated and in the conduit leading thereto, the container and the conduit being subjected to the full effects of the explosion. The present invention seeks, on the contrary, to ensure the protection of the very part of the installation where the explosion originates, RTI ID="0001.0266" WI="17" HE="3" LX="1058" LY=" 2272"> in particular in aircraft fuel tanks or other installations subject to explosions on board aircraft. It is needless to emphasize that in such applications it is essential to suppress an explosion when it is still in its initial stage, and to allow an explosion to develop its full force in any part of an airplane would almost certainly be a disaster. The present invention makes it possible to detect an incipient explosion and to neutralize it by distributing a liquid neutralizing agent in the zone of the explosion. The method that comprises the invention consists in detecting such an incipient explosion by the luminous phenomena which accompany it, that is to say by the initial flame of the explosion or by the source of ignition explosion, such as an incendiary bullet or spark, after which the neutralizing liquid is immediately dispensed at a speed of more than 45 meters per second.

The invention also comprises an apparatus for the implementation of such a method, comprising means for detecting the beginning of an explosion, a reservoir liable to break and containing the neutralizing agent and an explosive charge controlled by a electrical circuit. In accordance with the invention, the means for detecting the beginning of an explosion consist of at least one photoelectric cell intended to detect the luminous phenomena accompanying this beginning of the explosion and an amplifier of the flow rate of this cell capable of supplying the electrical circuit with a current for the ignition of the explosive charge.

The attached drawings show several embodiments of the apparatus comprising the invention given by way of example; these drawings also include explanatory graphs and diagrams.

fig. 1 is a graph showing the pressures on the ordinate and the time on the abscissa and showing the curve characteristic of the explosion of an optimum hydro-carbon-air mixture. fig. 2 is a plan view of an aircraft fuel tank, the upper wall of which has been removed to show the location of the photoelectric detectors and the distributors of an anti-explosive or neutralizing product.

fig. 3 is an end view of the reservoir shown in FIG. 2, from which the anterior wall has been removed.

fig. 4 is a diagram of a detector circuit of a first embodiment employing a photoelectric cell with amplification by secondary emission.

fig. 5 shows a detector circuit of a second embodiment using a simple photo-emissive cell associated with cold cathode lamps.

fig. 6 represents a variant whose circuit comprises a simple photo-emissive cell, associated with a thyratron.

fig. 7 shows another variant where the circuit comprises a photo-emissive cell associated with a hot cathode lamp and a magnetic relay.

fig. 8 is a graph showing, only approximately, the energy distribution in the visible spectrum, for typical explosion flames.

In the embodiments of the invention which will be described below, emphasis will be placed on the suppression of explosions in aircraft fuel tanks because such applications exhibit the greatest range of hydrocarbon-air mixtures. corresponding to the possibilities of explosion from very lean mixtures to very rich mixtures, in the presence of very unfavorable conditions of temperature, humidity, etc... Therefore, an explosion neutralization system which will give satisfaction in aircraft equipment will easily be able to suppress explosions in many other applications where the imposed conditions will be less severe.

The invention is not limited to applications in aviation but is also applicable to many other installations presenting risks of explosion, for example: diesel engine casings, installations for chemical catalytic reactions, electrostatic precipitation tanks, grinding and spraying, cyclones, filters and, in general, all other installations where finely divided explosive powders and dusts are handled.

fig. 1 shows a characteristic curve of the explosive rise in pressure in kilograms per square centimeter, P r, plotted on the ordinate, the times after ignition, t, measured in milliseconds being plotted on the abscissa; this curve shows an explosion test in a 200 liter tank with an optimum mixture of hydrocarbon and air. The pressure rise is characteristic of what happens when an optimum hydrocarbon-air mixture explodes.

It will be seen that the time required to reach an explosive pressure rise of 0.35 kg per ce is only about 50 milliseconds. We can demonstrate. that the rate of flame in the circumstances considered is practically independent of the volume of gas and it follows that the combustion of a mixture in a reduced volume is more rapidly completed than in a larger volume. Thus, the time required for an explosive pressure rise of <B>0.35</B> kg per cent in a volume of 20 liters containing an optimum mixture is only of the order of 23 milliseconds. Such volumes are common in practice, especially in the case of tanks almost full of fuel, and it is therefore clear that rapid explosions in small volumes spread so quickly that their detection must be carried out at a very low elevation. of pressure, if one wants to stop a beginning of explosion. The short delay before the pressure begins to build at a considerable rate is probably due either to the time required to build up the energy of the ignition source or to the time required to ignite the fuel.

To detect such explosions with the extreme rapidity necessary, photoelectric cells are employed, as has been said, which act extremely rapidly on illumination of the initial explosion flame or even of the ignition source itself. even, for example at illuminances of 5 lux or less.

However, rapid detection alone is not sufficient and must be combined with extremely rapid delivery of a neutralizing agent in liquid form. Therefore, the flame speed can reach a maximum of 3 meters per second; the neutralizing agent must be distributed at a speed which is not less than 45 meters per second, over a radius of at least 60 cm, i.e. in a period of time not exceeding 10 milliseconds .

For this reason, ordinary fire extinguishers and even extinguishers normally used in aviation for the rapid suppression of fires in engine cradles cannot be used. On the contrary, the reservoirs used containing the neutralizing agents must be designed in such a way that virtually the entire part of the reservoir receiving the liquid breaks into small fragments to distribute the neutralizing agent at the desired time, this reservoir also comprising means ensuring rant the projection of the liquid droplets at the necessary speed.

Therefore, the hemispherical shape can be used as for neutralizing agent reservoirs, which. case an explosive charge makes it possible to break the hemispherical tank and at the same time to project the neutralizing agent with the necessary rapidity. However, other shapes of neutralizing agent reservoirs capable of distributing this neutralizing agent with the same speed can be used. For example, spherical or other shaped neutralizing agent reservoirs may be employed provided that virtually the entire portion of the container containing the liquid breaks and the necessary velocity is imparted to the contents of the container.

Figs. 2 and 3 show sections taken through a typical aircraft fuel tank in which the fuel is contained in a flexible rubber container 10. This rubber container 10 is mounted in a frame adjacent to the fuselage envelope or the wing of the aircraft, which is usually reinforced at this point by tie rods, the inner chord of the reinforcement parts carrying a thin envelope forming part of the frame carrying the tank. Thus, the fuel is contained in the rubber container and the pressure is absorbed through the support by the wing or the fuselage itself. Alternatively, instead of the rubber container, the mentioned thin envelope can form the tank whose metal joints are suitably sealed.

As shown in fig. 2, a photoelectric detector 11a is mounted in an upper corner of the tank while a similar detector 11b is mounted in the diametrically opposite lower corner of the tank. The use of two detectors mounted in this way makes it possible, when the orientation of the airplane is such that one of the detectors is covered by a great thickness of fuel capable of attenuating in an appreciable way the illumination tom bant on the cell, to the other detector to be uncovered, or to be covered only to an acceptable extent.

The tanks. of neutralizing agent shown in FIGS. 2 and 3 are similar to those described in the Swiss patent N 326019 and com take two cylindrical tubes 12a and 12b fixed to the ends of the fuel tank 10 and extending over its entire length. The ignition circuit of the neutralizing agent reservoir is connected in parallel with two detectors 11a, 11b... and the connecting wires 13 are embedded in the rubber walls of the reservoir. Said tubes are particularly suitable for shallow reservoirs; for example, the reservoir shown may be 1 m 20 long by 90 cm wide and 30 cm deep. If spherical reservoirs of neutralizing agent are used, they can be placed, for example, as indicated by the dotted lines 14, although hemispherical reservoirs can also be mounted with their rear plate adjacent to one of the walls.

Although neutralizing agent tanks may contain a partly chemical and partly cooling neutralizing agent, such as methyl bromide or dibromodifluoromethane, it is also possible to suppress the explosion using a liquid that enriches the bare mixture beyond the explosive range. For this purpose, a liquid hydrocarbon, for example isopentane, can be employed. The great advantage of such an enriching liquid is that it has a much lower specific weight than that of neutralizing agents which act by cooling, the specific weight of the latter being three to three and a half times greater. In addition, the enriching liquids do not pose, with regard to the container, the problems of permeability and corrosion which arise for example with methyl bromide. On the other hand, when neutralizing, enriching agents are employed, there is a danger of re-ignition from a persistent source of ignition, particularly if the aircraft were to make a rapid descent which would result in a call for fire. air, bringing the mix back to the explosive range.

The boiling point of the neutralizing agent used is important as it will determine vapor pressure variations beyond the range of temperatures encountered in the particular application under consideration. Preferably, the neutralizing agent should be volatile enough to be effective at low temperature, although it does not have such a degree of vapor pressure at high temperature that it would require an excessively strong reservoir. For use in aviation, where the temperature may vary between -60 C and -I- 80 to 100() C, isopentane and dibromo- difluoromethane having a boiling point of about 260 under a pressure of 1 kg per cm2 are the most appropriate.

The dimensions of the neutralizing agent tank will determine the minimum spacing between the tanks which will give the required concentration of neutralizing agent based on the volume of the fuel tank. The smaller the containers used, the faster the distribution of the neutralizing agent will be, but this advantage will be obtained at the cost of a more complex installation due to the greater number of containers required.

As for the detector itself, we know that there are four basic types of photoelectric cells, namely photo-emissive cells (vacuum and gas), photovoltaic cells or film cells. shutdown, photoconductive cells and phototransistor or crystal junction cells.

Under aeronautical conditions, it is necessary to take into account maximum temperatures of the order of 1000 C and the only type of photoelectric cell which will operate satisfactorily at these high temperatures is the photo-emissive cell which is also the cell most sensitive to the visible part of the spectrum. Although it may be possible to insulate other types of photocells from high temperatures, this will generally be a complication to be avoided.

There is a particularly sensitive type of photo-emissive cell in which an electronic multiplier is incorporated and which will be called hereafter photoelectric cell with amplification by secondary emission or, more briefly, photoelectric cell with secondary emission. For the kind of application examined here, the secondary emission photoelectric cells can be of an order of sensitivity 100 times greater than that of a simple photoelectric cell and have a sensitivity greater than 1 microampere per lux but; of one. on the other hand, they have the disadvantage of requiring a source of high potential for the secondary anodes of the multiplier. The voltage required is usually on the order of 70 to 100 volts per multiplier stage, which with today's multipliers means a total of 700 volts or more.

Commercially, there are no suitable photoelectric cells which have sufficient power output to ignite the explosive charges of the neutralizing agent reservoirs. In the case of photoelectric cells, including secondary emission photoelectric cells, the amplification of the output current is advantageously carried out electronically to ensure the almost instantaneous ignition of the neutralizing agents during detection. The amplifier may consist simply of one or more cold cathode tubes, which have no fragile filaments and are of robust construction.

The first embodiment (Fig. 4) comprises a secondary emission photoelectric cell 20, represented by its capped and pinned end connections, pin 0 forming the connection with the photocathode, pins 1 to 9 forming the connections with the multiplier secondary cathodes and pin 10 forming the connection with the collector anode. Pin 0 is connected to one end of the secondary winding of a step-up transformer 21, the primary winding of which is connected to an alternating current power source 22. This is advantageously a voltage supply. high voltage and high frequency for aircraft which is normally 115 volts, 400 periods, and which is transformed by the transformer 21 into a voltage of the order of 600 volts. The other end of the secondary winding is connected via a 1.5 megohm resistor to ground and pin 8 and via a 7 picofarad capacitor to pin 9.

The collector anode is connected via pin 10 and a 1 megohm resistor to the positive terminal of a 125 volt DC power source, the negative terminal of which is connected to ground. The current amplifier lamp is in the form of a cold cathode triode 23, the control grid of which is connected to pin 10. The cathode of the triode 23 is connected to earth by the circuits, in parallel, of ignition of the explosive charges 24; in a variant, these circuits are in series. A 4 microfarad filter capacitor is connected across the 125 volt power source. A resistor comprising two elements of 7 and 11 megohms forms a voltage divider between the terminals of the 125 volt power source and biases the control grid of the triode 23, so that this lamp is normally off. driver. When the photocathode is illuminated, the sharply increasing current between the last second cathode and the collector anode virtually shorts the 7 megohm resistor and the increased voltage applied to the control electrode causes the lamp 23 to conduct. , the subsequent discharge of the capacitor by 4 microfarads igniting the explosive charges. The 1.5 megohm resistor and the 7 picofarad capacitor are used to compensate for the troublesome internal capacitance of the secondary emission photoelectric cell which, under the considered supply frequency, tends to reduce the impedance in the upper branch considerably. of the voltage divider. The secondary emission photocell is equipped with a blue sensitive photocathode. Although greater sensitivity can be obtained above the maximum of the explosive range, if the cathode is red sensitive, as will be explained later, this cell is much more sensitive to explosive mixtures over the entire range than a simple red-sensitive gas cell.

The danger of having a source of the order of 1000 volts inside or near the fuel tank is reduced by mounting the transformer in close proximity to the cell and isolating the transformer, this cell and connections between them by embedding them in a plastic block, the light-sensitive part of the cell emerging from this. block. This resin envelope also serves to prevent the access of fuel and water vapour. Although cold cathode lamps should be preferred for amplifying cell output current because of the absence of a heated filament and their greater resistance, variations include normal electronic tube amplifiers. However, such amplifiers cannot increase the effective sensitivity of the photoelectric cells, since, for very small illuminations, the rate is of the order of magnitude of the small occasional variations in the rate of the cell which occur even when the cell is in complete darkness. This occasional flow lowers the limit of illumination that can be detected by means of photoelectric cells. On the other hand, the use of electronic multiplier stages as in the secondary emission photoelectric cell gives amplification without increasing the dark current of the cell.

When a very high sensitivity of a secondary emission photoelectric cell is not necessary, a variant comprising a simple photo-emissive cell, sensitive to red, for example, is used. This photoelectric cell can be a vacuum cell, but preferably a gas cell is used because such cells have a higher sensitivity than an equivalent vacuum cell.

The embodiment of FIG. 5 comprises a photoelectric cell 30 connected to the terminals of the upper branch of a voltage divider formed by two resistors of 15 and 18 megohms respectively and, as in the case of FIG. 4, the illumination falling on cell 30 produces a voltage rise on the control electrode of a cold cathode triode 31 which is thus made conductive. This triode 31 is in turn connected in parallel with the upper branch of a voltage divider comprising resistors of 140 and 190 kilo-ohms, the junction point between the resistors being connected to the control grid of a tetrode at Cold cathode 32. This tetrode is connected in parallel with a 250 microfarad filter capacitor and the ignition circuit for the explosive charges 33 of the neutralizing agent distributors is interposed in the triode cathode circuit. The value of the second voltage divider is also chosen so that, normally, tetrode 32 is non-conductive, but the voltage rise occurring when triode 31 becomes conductive makes tetrode 32 conductive and the 250 microfarad capacitor is discharged by the tetrode 32 to ignite the explosive charges.

fig. 6 shows the circuit of a variant powered by a source 41 of alternating current, without rectification, the photoelectric cell 40 being provided to control a thyratron 42 and ignite the explosive charges 43 connected in series, when it becomes conductive. Thyratrons have the disadvantage of having a heated filament but are capable of passing a stronger current than a cold cathode lamp of the same dimensions.

fig. 7 shows a variant whose circuit comprises a photoelectric cell and a hot cathode triode. The control grid of the triode 51 is connected to the anode of a photoelectric cell 50 whose cathode is connected to a source of negative bias. When the photoelectric cell 50 is not illuminated, the grid has no influence, so that the tube 51 passes a strong anode current which excites an electromagnetic relay 52. The relay 52 maintains a contact 52a which is interposed in the ignition circuit, in the open position. The illumination falling on the cell 50 will cause a negative polarization of the grid of the lamp 51 and, consequently, the breaking of the anode current; the relay 52, being no longer energized, will close the contact 52a in the ignition circuit of the explosive charges 53. The effective amplification which it is possible to obtain by means of this circuit is very great and the device is comparable in sensitivity to that shown in FIG. 4. However, it suffers from the disadvantage of requiring a heated filament and also from the fact that the slightest fault in the circuit causes the relay contacts to close and the neutralizing agent distributors to start; therefore, it is not desirable to use this device in aviation. However, for other applications where these disadvantages are acceptable, it constitutes an advantageous variant of the embodiments whose circuit comprises a photoelectric cell with secondary emission.

The energy emitted by a hydrocarbon explosion flame is not limited to a certain wavelength or a certain range of wavelengths and the energy is not the same for each range. The spectral energy distribution results from the combination of separate characteristics of each component of the flame, these components varying with the type of flame.

The most important factor is the presence or absence of free carbon particles, which occur especially in mixtures which are richer than they should be, and their presence greatly increases the radiant energy. from the wavebands of the visible spectrum and into the infrared. However, the types of photoelectric cells which have the maximum sensitivity in the infrared region are those which show reduced sensitivity at temperatures above 200 C, and for the application to aviation it is necessary to employ photoelectric cells with maximum sensitivity in the visible spectrum.

At the end of the range of mixtures corresponding to lean mixtures, there is no deposit of free carbon particles and the spectrum of the explosion flame is then simply due to the gases. This results in a very large reduction in total radiation energy compared to what happens with rich mixtures, as seen in fig. <B>8</B>, which shows the approximate energy (E) distribution in the visible spectrum, with curve A corresponding to very rich mixtures and curve B to lean mixtures. The general shape of the curve is due to black body radiation from carbon particles, while the maxima for wavelengths of approximately 3000, 4000 and 5000 Angstrom units are due to radiation from combustion gases.

It will be clear from examination of fig. 8 that, while the color of the flame is normally red in rich mixtures, it becomes relatively less red as the mixture becomes leaner, the increase in the blue of the flame being accompanied by a weakening of its intensity bright.

Although secondary emission photocells are sensitive enough to react to reduced brightness produced by the leaner mixtures of the mixture range, if simple emissive cells are employed, other embodiments include two cells connected in parallel, one cell being sensitive to blue while the other is sensitive to red. The circuits of these embodiments are analogous, respectively, to the circuits of FIGS. 5 to 7. The use of two cells connected in parallel naturally improves the sensitivity of the device over the major part of the range of mixtures, certain variants include more than two cells of a given sensitivity. It is possible to obtain an appreciable increase in the radiation from a burning gas by adding very small quantities of certain substances to the fuel in the tank. Such substances must vaporize at the same time as the fuel and be able to be stored in a tank. They must also not cause damage to the engine or the pipes. An example of such substances is iron pentacarbonyl.

As is well known, the illumination decreases according to the square of the distance from the illuminated object to the source. However, in a tank with bleached walls,. the reflection of the radiation coming from the inside of the tank increases the sensitivity of the detection and moreover the said inverse square law is no longer valid. It is therefore possible to obtain an improvement of the order of a multiple of 4 by whitening the walls of the tank, whereas with a shiny metal inside the tank the improvement is much lower. Additionally, optical amplification means may be employed to increase the illumination falling on the photoelectric cell.

Other embodiments include photovoltaic or barrier layer cells. They are not suitable for use in aviation, given the loss of sensitivity at the relatively high temperatures at which these cells. may be exposed, in addition to the fact that difficulties are experienced in amplifying the throughput of the photoelectric cell due to the low load impedance which must be employed. This type of cell has the advantage that it is self-generating and does not require an external power supply. A typical output current flowing into 1000 ohms is about 4 microamps, which is sufficient to directly operate a relay operating at high speed, for example a biased relay. When the current is not sufficient for the safe operation of a relay, because the latter is subjected to vibrations, an embodiment is used comprising means for increasing the output current. These means consist of a mosaic of cells connected in series and a magnetic amplifier which is more suitable than an amplifier with electronic lamps. Although such amplifiers are subject to response delays on the order of a few milliseconds depending on the frequency of the AC supply and the power boost required, the use of such an embodiment is however very practical where such delays are tolerable and where the maximum temperature is not such that it can excessively reduce the sensitivity of the cell.

Individual mosaic cells have different spectral sensitivities for the reasons already stated.

A variant of the embodiments described with barrier layer cells comprises a transistor-type crystal amplifier to amplify the output current of the barrier layer cell.

Other embodiments include transistors as sensor elements. It has been observed, in fact, that certain forms of transistors possess photoelectric properties. One or more transistors having such properties can be used in combination with amplifying transistors or other amplifying devices. A very compact detector can be obtained by using similar transistors associated with an electromagnetic relay operating at high speed arranged so as to close the ignition circuit.

The sensitivity and stability of phototransistors are not comparable to those of secondary emission photoelectric cells, for the moment.

Finally, a last form of execution comprises, as detectors, photoconductive cells whose maximum sensitivity extends well inside the infrared zone of the spectrum where, as indicated previously, the major part of explosion flame donkey radiation. For aviation, this type of cell is not suitable, given the limitations imposed by the temperature. In addition, an alternating current amplifier must be employed to amplify the output current, but since the latter is in the form of a direct current signal, it is necessary to make arrangements to periodically modulate or interrupt the radiation falling on the cell to be able to amplify the pulsating output current of the cell. The flickering of the explosion flames is not suffi cient to remove the need for such periodic interruption, at least at the stages where detection is necessary for the neutralization of the explosions.

For applications where these disadvantages are bearable and particularly where the radiation is mainly infrared, the embodiment with photoconductive cells is entirely appropriate.

The neutralization devices described are also applicable in installations presenting other explosion risks, for example in the case of dusts and powders. The choice of detector must be determined after examination of the various conditions to be considered, such as the maximum ambient temperature and the spectral distribution of the illumination resulting from an incipient explosion in the explosive atmosphere to be expected. Similarly, the choice of neutralizing agent and the concentration and rate of delivery required will depend on the conditions associated with the particular hazard being considered.

Claims (13)

CLAIMS 1. A method of detecting an incipient explosion and neutralizing it by distributing a neutralizing liquid agent in the area of the explosion, characterized in that the incipient explosion is detected by the luminous phenomena which accompany it and which, following this detection, the neutralizing liquid is distributed at a speed of more than 45 m. per second. II. Apparatus for carrying out the method of detection and neutralization according to claim I, comprising means for detecting the beginning of an explosion and a reservoir capable of being broken and containing the neutralizing agent as well as a controlled explosive charge by an electrical ignition circuit, characterized in that the means for detecting the beginning of an explosion consist of at least one photoelectric cell intended to detect the light phenomena accompanying this beginning of the explosion and an amplifier of the current produced by this cell, capable of supplying said electrical circuit with a current for igniting the explosive charge. SUB-CLAIMS 1. Process according to claim 1, characterized in that the neutralization of the explosions is produced by an enrichment of the explosive mixture. . 2. Apparatus according to claim II, characterized in that the photoelectric cell is a photo-emissive cell. 3. Apparatus according to claim II, characterized in that the photoelectric cell is a secondary emission photoelectric cell. 4. Apparatus according to claim II, characterized in that the current amplifier comprises a cold cathode lamp, a voltage divider intended to polarize a control electrode of the lamp and to make the lamp normally non-conductive and a filter capacitor connected to an electrical source and mounted in series with said lamp and said ignition circuit, the output of the photoelectric cell being interposed between the anode and the control electrode of said lamp, in such a way so that the lamp is made conductive when the photoelectric cell is exposed to the aforesaid illumination and that the capacitor is then discharged by the ignition circuit and ignites the explosive charge. 5. Apparatus according to claim II and sub-claim 3, characterized in that the sensitivity of the secondary emission photoelectric cell is greater than one micro-ampere per lux. 6. Apparatus according to claim II and sub-claim 3, wherein the high voltages required for the multiplying secondary cathodes of the cell come from a transformer which is connected, by high voltage connections, inside the envelope of the secondary emission photoelectric cell, characterized in that said transformer and the high voltage connections are embedded in a block of insulating plastic material. 7. Apparatus according to claim II and sub-claim 3, characterized in that the secondary multiplier cathodes of the secondary emission photoelectric cell are supplied directly with alternating current. 8. Apparatus according to claim II, ca acterized in that it is arranged to react to an illumination of less than about 5 lux. 9. Apparatus according to claim II, characterized in that it comprises several photoelectric cells connected in parallel and each having a different spectral sensitivity. 10. Apparatus according to claim II, characterized in that the photoelectric cell is a photovoltaic cell and that its output circuit is connected to an electromagnetic relay. 1l. Apparatus according to claim <B>II</B> intended to be installed in an aircraft fuel tank, characterized in that it comprises two photoelectric cells intended to be mounted one in an upper corner of the fuel tank and the other in the diametrically opposite bottom corner of the tank. 12. Apparatus according to claim II, characterized in that the neutralizing agent is a combustible liquid hydrocarbon. 13. Apparatus according to claim II, characterized in that the boiling point of the neutralizing agent under a pressure of 1 kLy/cm2 is approximately 260 C.